NEWSLETTER OF THE DEPARTMENT OF COMPUTER SCIENCE AT COLUMBIA UNIVERSITY VOL.10 NO.1 SPRING 2014

CS@CU SPRING 2014 1

CUCS Welcomes Two New Faculty

Allison Bishop lewko received her Ph.D. in Computer Science from the University of Texas at Austin in 2012. She also received a Certificate of Advanced Study in Mathematics from the University of Cam-bridge in 2007 and her Bachelor’s degree in Mathematics from Princeton University in 2006. Her research interests include cryptography, complexity theory, distributed computing, harmonic analysis, and combinatorics. Prior to joining Columbia, she was a postdoctoral researcher at Micro-soft Research New England.

CUCS: How did you first become interested in computer science?

I took a somewhat long and indirect route to computer science. I began college as an english major because I wanted to do creative writing, but then I

started taking some pure math-ematics courses and discovered that math can be creative in ways I hadn’t realized. After starting my graduate studies in math, I went looking for poten-tial applications of mathematics to motivate my research. I found that computer science offers a great mix of deep theoretical challenges and the promise of practical impact. So midway through my doctoral studies I became a computer scientist.

What attracted you most to a faculty position at Columbia?

Columbia is home to many truly stellar researchers that I expect to collaborate with over the coming years. I am also deeply impressed by the excellent students at Columbia and feel very lucky to have the opportunity to work with them.(continued on next page)

“Allison lewko is using math-ematics to revolutionize the field of modern cryptography. Her work on functional encryption provides a fundamentally new way to think about the classical concept of encryption.”

– TAL MALKIN

Daniel Hsu and Allison Lewko

Assistant Professor of Computer Science Allison Lewko

CS@CU SPRING 2014 3

The old adage, “a picture is worth a thousand words,” is particularly fitting for the creators of WordsEye.

The text-to-image software startup, co-founded by com-puter science PhD candidates Bob Coyne and Daniel Bauer, recently won the grand prize—a $100,000 award—at the Fourth Annual New York Business Plan Competition. WordsEye bested more than 430 teams from nearly 60 colleges and universi-ties statewide in the competi-tion, sponsored by the State University of New York’s College of Nanoscale Science and Engineering (CNSE) to support high-tech education and entre-preneurship amid the state’s thriving technology scene.

The WordsEye team, which also includes Columbia Business School student Neelam Brar, intends to use the award money to further develop its propri-etary software, which the team describes as “Twitter meets Instagram.” WordsEye’s technol-ogy, geared toward social media users, provides a new medium for visual expression, allowing users to create 3D scenes or im-ages simply by text descriptions.

“The goal of WordsEye is to help people express themselves in visuals as quickly and easily as they can with language,” says Coyne. “It took a while for the news to sink in. Our next step is to get financing to actually launch the business.”

The team’s momentum is in overdrive, having been in the works over the past several years. Coyne says they are already talking to several potential investors, plan to participate in more pitch events, and aim to beta test the software this summer.

The team has already gained in-terest from the educational sec-tor. The software can be used for vocabulary building, literacy, special needs instruction, cyber arts, and storytelling. “We tested WordsEye with sixth-grade students in an enrichment program a couple of summers ago at Harlem Educational Activities Fund,” says Coyne, “and found that students using it improved more significantly in their tested literacy skills as compared with a control group.” Down the road, the team will look into applying WordsEye to 3D games.

Julia hirschberg, Professor and Department Chair of Com-puter Science, is Coyne’s faculty adviser and has been involved early on in the making of Word-sEye. With Columbia Research Scientist owen Rambow, and Richard Sproat (now at Google), Hirschberg helped secure fund-ing from the National Science Foundation to conduct additional research on WordsEye and get the product where it is today.

“Now we’re at the stage where WordsEye can be made widely available, through social media,” says Hirschberg. About the sig-nificance of the software, she remarks, “Being able to draw beautiful pictures using regular language seems like magic to WordsEye users, even though it’s based on a great deal of very sophisticated language processing technology.”

Coyne got the idea for Words-Eye after his then two-year-old son woke him in the middle of the night and he was unable to get back to sleep. He had been developing a commercial 3D graphics system at the time and felt frustrated that he never had the time to actually use the system himself; the system was flawed in that it was too laborious and time consuming.

“I wanted to much more easily translate my own artistic ideas into actual scenes. It occurred to me that language would let me do that,” recalls Coyne. “By using language you give up some amount of control over details but you gain in expres-sive power.”

WordsEye’s primary focus is social media, as a tool, adds Coyne, “to let everyone create pictures that they can share. With WordsEye, people won’t have to spend months learning complex tools and hours using those tools in order to create beautiful rendered 3D scenes.”

By Melanie A. Farmer

Student-led Social Media Startup Wins Grand Prize in Statewide Competition

What are some of your ongoing research projects and what motivated you to focus on these directions?

I am currently working on two main research directions. One direction is to extend the reach of provable security guarantees for cryptographic systems, while the other is to develop error-resilient algorithms and lower bounds for distributed computation and communica-tion tasks. These are problems that have many potential applications, but our current understanding of the inherent tradeoffs between the com-peting goals of efficiency and security is quite lacking.

Do you have a method or approach that you like to use in your research?

I do what probably most theoreticians do—I ask a lot of questions and I generate a lot of bad ideas! I usually start with many ideas that individu-ally don’t work, try to figure out how and why each fails, and extract insight from this that can be mined for solutions.

How do you pick your research topics and what do you believe makes a good research question?

I am typically drawn to prob-lems that populate the boundar-ies between what is theoreti-cally possible and impossible, and I seek to understand what are the core features of a problem that tip the balance between a potentially practical solution and an impossibility result. I think a good research question is one that isolates the most basic gap in current understanding and suggests a path towards further progress.

Cover Story (continued)

2 CS@CU SPRING 2014

Student Awards

DAniel hsu received his Ph.D. in Computer Science in 2010 from the Department of Com-puter Science and Engineering at UC San Diego, where he was advised by Sanjoy Dasgupta. He was a postdoc at Microsoft Research New England from 2011 to 2013; before that, he was a postdoc with the Depart-ment of Statistics at Rutgers University and the Department of Statistics at the University of Pennsylvania from 2010 to 2011, supervised by Tong Zhang and Sham M. Kakade. Daniel received his B.S. in Computer Science and Engineering in 2004 from the Department of Electrical Engineering and Computer Sciences at UC Berkeley. His research interests are in algorithmic statistics and machine learning.

What attracted you most to a faculty position at Columbia?

The people here are enormously talented and accomplished, and also ever friendly and generous. This combination makes for a wonderful academic home.

What are some of your ongoing research projects and what motivated you to focus on these directions?

One project aims to develop machine learning algorithms that preserve privacy in a rigor-

ous sense. This is important because many applications of machine learning rely on personal and sensitive data about individuals (e.g., genome sequences, financial histories). Another project concerns machine learning in interactive settings, where the data and feedback that are collected depend on specific choices made by the algorithm. These settings are ubiquitous, yet require techniques which are fundamentally different from more “classical” machine learn-ing approaches. A third project looks at algorithms for deriving new representations of data useful for machine learning. This is part of a broader goal of making machine learning more widely usable and automatic.

How do you pick your research topics and what do you believe makes a good research question?

Topics of research are, ideally, necessary and sufficient for progress in some grander scheme. I think many people (myself included) are also attracted by other aspects, such as aesthetics and practical significance. One strategy is to pick a problem or question such that whatever the eventual solution or answer is, it will still be interesting to you.

Where do you see your field progressing over the next decade?

At the core, there is still much effort put towards making machine learning a more au-tomatic, reliable, and scalable technology. This is necessary for the field to mature, and I think the next decade will see significant advances in these foundational issues.

If you were advising students at the undergraduate level who wanted to do what you’re doing now, how would you suggest they go about it?

A lot of practice in both algorith-mic design and mathematical analysis goes a long way in machine learning.

This WordsEye image was generated from the following text: “The orange cat is on the chair. The chair is white. The ground has a design texture. The ground is shiny. The huge yellow illuminator is 7 feet above the cat.”

“Daniel hsu pioneered novel approaches to learning statistical models from data by using spectral techniques and the method-of-moments. While previous approaches got stuck at suboptimal solutions, Hsu’s work guarantees globally optimal models in a computationally efficient manner.”

– TONY JEBARA

Assistant Professor of Computer Science Daniel Hsu

4 CS@CU SPRING 2014

Feature News

CS@CU SPRING 2014 5

A long-awaited project has come to life for shree nayar: Bigshot, a kit that features a build-it-yourself digital camera he has designed to serve not only as a creative tool but also as a medium for education. More than four years in the making, Bigshot is the culmina-tion of a dream for Nayar, who was inducted into the National Academy of Engineering in 2008 as well as the American Academy of Arts and Sciences in 2011 for his pioneering contri-butions to computer vision and computational imaging.

“The camera is central to what we do,” says Nayar, T.C. Chang Professor of Computer Sci-ence at Columbia Engineering. “It’s more than just a device that captures images in time, memories in pixels. It allows us to express ourselves and to communicate with each other in powerful ways—it has evolved into a technology with great social relevance.”

Bigshot, which Nayar developed for anyone interested in how cameras work (“kids” from 8 to 108), includes features like a hand-cranked power generator to supplement the rechargeable battery, and a “Swiss Army” lens with a wheel you can ro-tate to switch between regular, panoramic, and stereo or 3D modes. The kit is accompanied by an in-depth website that includes an interactive textbook with engaging demos of the sci-ence and engineering concepts related to the camera—optics,

mechanics, electronics, and image processing.

The project was inspired by Born into Brothels, a documen-tary about desperately poor children living in the red light district of Calcutta who were given cameras by one of the di-rectors. Nayar had already been thinking about how to leverage the appeal of the camera to make a broader social impact when he saw the film, and so he embarked on Bigshot as a side project to his research.

“As the kids learned photogra-phy, they began to look at their world with new eyes,” Nayar observes. “The film reaffirmed my belief that the camera is unique as a technology in its ability to inspire creativity.”

He notes that building Bigshot component by component ex-poses children to a wide range of science concepts, and they learn by doing: “Once they have a fully functioning digital camera on their hands, they get to enter the world of photography, storytelling, and documenta-tion. The inherent nature of the camera enables us to juxtapose the sciences and the arts within a single learning experience.” Younger children benefit from working with their parents and teachers on the project; older teens and camera enthusiasts enjoy exploring the camera and website on their own.

At Columbia, Nayar worked with independent contractors to build about a dozen prototypes

of the camera. He and a team led by graduate students Guru krishnan and Brian smith developed the first version of the educational website. They field-tested the prototypes and the website over a six-month period with hundreds of kids and their teachers in four cities—New York, Bangalore, Vung Tau, and Tokyo.

The camera kit was so well-received that Nayar founded Kimera (short for kid camera) in 2010 as a for-profit social venture to manufacture and distribute the Bigshot kits. His goal is to use some of Kimera’s royalties to donate cameras to kids in highly underprivi-leged communities through a program he calls Bigshots for Good. He is hoping children will share their photos online, and on the Bigshot website. And he intends to partner with educational nonprofits and after-school programs—he already has one partnership inked with the Center for Arts Education in New York, to bring Bigshot to underserved students.

“To me Bigshot is not just a kit, or a digital camera,” Nayar says. “It’s an experience that I hope will pique curiosity. One kid may be drawn to the science behind it and another may be inspired by the art of photography. Both would be equally good outcomes in my mind.”

By Holly Evarts

Bigshot Camera Does Good

Nayar aims to help children combine creativity with science using Bigshot

PhoT

o of

ShR

EE N

AyAR

: JEf

fREy

SCh

IfM

AN

Shree Nayar hopes children will learn from first building his camera and then sharing photos

Faculty News & Awards

Allison lewko, Assistant Professor of Computer Science, has been named to the annual Forbes 30 Under 30 list in the Science & Healthcare category for her work on designing en-cryption algorithms and keeping data secure. She joins an exciting group of, as the Forbes editors note, “young disruptors, innova-tors, and entrepreneurs who are impatient to change the world.”

“I’m very proud to be recognized as part of such a talented group,” says Lewko, who is focused on developing new techniques to achieve flexible data sharing and enhanced security in the

cloud, for example, how to keep secrets on devices like cloud-connected cell phones that are frequently leaking information.

“We call this functional encryp-tion, a way to enable only the users with specific attributes that match the encryption policy of the data to decrypt the docu-ment,” she explains. She sees functional encryption as a new method for obtaining strong security guarantees while still allowing a variety of uses of sensitive data. “In previous models of encryption, encrypted data would be safe but not very usable,” she adds. “We are

modernizing these concepts and creating tools that can scale with the size and variety of data and users in current systems.”

Before coming to Columbia Engineering in fall 2013, Lewko was a postdoctoral researcher at Microsoft Research New England, where she designed error-tolerant algorithms. She is a faculty affiliate of Columbia’s Institute for Data Sciences and Engineering and the Institute’s Cybersecurity Center.

By Holly Evarts

Forbes 30 Under 30 Recognizes Rising Star Allison Lewko

Assistant Professor Allison Lewko

Allison Lewko, Assistant Professor of Computer Science, has been named to the annual forbes 30 Under 30 list in the Science & healthcare category

CS@CU SPRING 2014 76 CS@CU SPRING 2014

Faculty News & Awards (continued)

Roxana Geambasu, Assistant Professor of Computer Science, has won a National Science Foundation (NSF) CAREER Award for her proposal to create new data protection abstrac-tions for modern operating sys-tems. The five-year, $499,000 grant, the most prestigious NSF award to recognize outstand-ing junior faculty, will help fund her project, “New Abstractions for Responsible Sensitive Data Management in Modern Oper-ating Systems.”

“This is a great honor,” says Geambasu, who is one of three faculty members directing the Software Systems Lab and a member of the Institute for Data Sciences and Engineer-ing’s Cybersecurity Center. “We are looking at creating new, con-venient protection mechanisms that will promote a responsible approach to data management, in which users can manage their data carefully and minimize their exposure to attacks.”

Her research is focused on en-suring data security and privacy in an era of cloud computing and ubiquitous mobile devices. She notes that these technolo-gies, which billions of users rely upon to access and host sensi-tive data—such as emails, docu-ments, and financial informa-tion, have become easy targets for theft, espionage, hacking, and legal attacks. “And, despite the threats,” she observes, “today’s data management practices are looser and more irresponsible than ever.”

These days, she points out, the mobile devices we carry with us everywhere are packed with confidential information under operating systems that never securely erase data. At the other end, cloud services not only accumulate endless logs of user activity, such as searches, site visits, and locations, but also keep them for extended periods of time, mine them for business value, and at times share them with others—all without the user’s knowledge or control.

“This has become an untenable situation,” Geambasu says. So she is working to identify the security risks inherent in current mobile and web technology and designs, and constructing systems to address those prob-lems. Her NSF CAREER-funded proposal is focused on look-ing at data storage in modern operating systems, including Android and iOS, and develop-ing fine-grained data protection.

“Ensuring mobile and cloud data security demands new mechanisms and abstractions to facilitate programmers’ work; ensuring privacy demands new levels of transparency that empower users to protect themselves,” Geambasu says. “My research addresses these challenging problems and antici-pates those to come with the archiving of vast quantities of sensitive data.”

By Holly Evarts

Roxana Geambasu Wins NSF CAREER Award

Assistant Professor Roxana Geambasu

Roxana Geambasu, Assistant Professor of Computer Science, has won a National Science foundation (NSf) CAREER Award for her proposal to create new data protection abstractions for modern operating systems

Augustin Chaintreau Wins CAREER Award, Rising Star Award

Augustin Chaintreau, As-sistant Professor of Computer Science, is on a roll. He recently won a prestigious National Science Foundation (NSF) CA-REER Award for his Banalytics research project that demon-strates how information about a consumer’s behavior, however mundane, can be used with tremendous values and risks to the individual. One of the NSF’s highest honors for exceptional junior faculty, the CAREER Award will support his work with a $500,000 five-year grant. In addition, Chaintreau has been awarded the distinguished 2013 ACM (Association for Comput-ing Machinery) SIGMETRICS Rising Star Researcher Award for his significant contributions to the analysis of emerging distributed digital and social networking systems.

”This is truly a double honor,” says Chaintreau, who also is a member of the Institute for Data Sciences and Engineer-

ing’s New Media Center. “My earlier work has now received its highest recognition. And being awarded this prominent NSF CAREER grant will ac-celerate my lab’s research at a time when reconciling the rapid progress in big data with per-sonal privacy has never been so critical. This is so gratifying!”

With more and more companies using customer data as a central component of their business plans, the world is becom-ing ever more dependent on researchers and policy makers who can access individuals’ personal data. “Today, customer data is what makes a company profitable. Tomorrow, sharing our personal data will offer ways to improve our lives and ad-vance society. So how,” Chain-treau asks, “do we reconcile this progress in big data with privacy? This is a problem that’s getting worse every day, espe-cially now that personal data and its value for national security has become such a controversial—and important—issue.”

Modern-day analytics, defined by Chaintreau as the “science of identifying and exploiting individual types and trends,” currently run behind closed doors—you cannot control how companies like Facebook or Google use and market user data. “The first step,” he notes, “is to prove you can perform analytics with data transparen-cy, where personal information is controlled by individuals who

can then decide what they want to share—or not.” His research is targeted at finding alterna-tives that will help people man-age their personal data in an easier, more transparent way. Data transparency, he explains, would make big data socially more efficient—“but once you prove it’s technically feasible, you face the real question of how can we overcome adoption barriers so that the web as we know it will evolve towards bet-ter transparency?”

“Privacy and economics are re-ally intertwined,” he continues. “Users and companies will not work together if the solution handles one well at the expense of the other. Enabling users to effectively own their data and trade them with companies raises all kinds of technical, eco-nomic, and societal challenges. Answering that question will decide if the mobile web we inhabit in 10 years is a hospi-table and prosperous place, a means of unprecedented surveillance, or a vast collection of fiefs guarded by metal doors. Researchers, app developers, and regulators, all are making progress, but nobody has yet gotten the path to a transparent web completely right. That’s what we are trying to do.”

Along with building behavioral analytics that put users in con-trol, Chaintreau plans to create a new course for both engineer-ing and journalism students on personal data management.

“Some of the most innovative work measuring how compa-nies handle personal data come not from the conferences I attend, but from the newspa-pers and blogs we all read,” Chaintreau says. “At Columbia, I am very fortunate because we can bring together the best of computer science and journalism. In fact, we are now working on launching a pilot class in particular for the dual MS degree program on how to inform society about the risks and values of personal data.”

It is getting increasingly more important for individuals to un-derstand the value of their data.

“In 10 years all students should understand the value of information in our evolving society,” he adds. “This is really a civic responsibility—we need good technologists and, just as importantly, good journalists who are truly tomorrow’s data first-responders to improve how citizens make critical choices. Surrendering our data for some free services, for instance, is not a good option for critical areas like health and energy. I hope that my research will shape a different future for the mobile web.”

By Holly Evarts

Assistant Professor Augustin Chaintreau

Chaintreau’s Banalytics project demonstrates how information about a consumer’s behavior, however mundane, can be used with tremendous values and risks to the individual.

CS@CU SPRING 2014 98 CS@CU SPRING 2014

Faculty News & Awards (continued)

Rocco servedio, Associate Professor of Computer Science, and David Vallancourt, Senior Lecturer in Electrical Engi-neering, have been awarded Columbia’s Presidential Teaching Award, a top teaching honor given by the University to just five recipients each year at Commencement. Established in 1996, the presidential awards recognize the best of Colum-bia’s teachers “for the influence they have on the development of their students and their part in maintaining the University’s longstanding reputation for educational excellence.”

Both Servedio and Vallancourt joined the faculty procession at the University Commencement Ceremony held May 22, 2013.

“We are extremely proud that Rocco and David have been honored this year by the University for their outstanding teaching,” said Donald Goldfarb, interim Dean of Columbia Engi-neering. “SEAS has two of the five awardees this year, which really underscores the great strength of our faculty and their commitment to teaching.”

Servedio and Vallancourt were surprised to hear the news.

“I was thrilled,” says Servedio. “There are so many tremendous-ly dedicated and skilled teachers at Columbia, all across the dif-ferent schools of the University, that it was really very meaningful to get this recognition.”

As teachers, both aim to keep it interesting. Their teaching styles differ but they agree it is key to be able to engage the students in the material and help them really understand it.

Vallancourt, who teaches the first-year engineering course in electrical engineering, has been credited for keeping the lectures fresh, fun, and full of interesting demonstrations. “One of my main tools is making connections between course material and common devices with which students are familiar,” says Vallancourt, who has in the past dissected everything from smartphones to electric guitars in class. “This gives me the opportunity to tear things apart on ‘company time,’ and what engineer wouldn’t enjoy that?”

Servedio aims for a relaxed classroom style “with a heavy emphasis on making sure that students first have a solid intuitive and conceptual grasp of the material before getting into mathematical details.”

Servedio teaches an intro-ductory course in computa-tional learning theory and it is important to him to try to explain the material clearly and convincingly so that students understand why it is interesting and worthwhile. “A lecture is a lot like a story, and if a story doesn’t make sense on a ‘gut level’ then there is no way a listener is going to care enough to absorb the details.”

A leading computational theo-rist, Servedio joined Columbia Engineering from Harvard where he was an NSF Mathe-matical Sciences postdoc in the University’s Division of Engi-neering and Applied Sciences. He studied at Harvard, earning his AB in Mathematics and both his MS and PhD in Computer Science. Servedio’s research focuses mainly on computa-tional learning theory as well as computational complexity theory, cryptography, and quan-tum computation.

Vallancourt, who is also an alumnus of Columbia Engineer-ing, having earned his bach-elor’s, master’s, and PhD here all in electrical engineering, joined the School in 1987 as an assistant professor. In 1992, he left Columbia for AT&T Bell Labs (later Lucent Technologies) where he served on the techni-cal staff until 2000. Before re-turning to Morningside in 2005 as a senior lecturer in Circuits and Systems in the Department of Electrical Engineering, he worked at Texas Instruments and Vitesse Semiconductor.

This is not the first time Val-lancourt has been honored for exemplary teaching. In 2007, he received a Distinguished Faculty Teaching Award, which is pre-sented annually on Class Day by the Columbia Engineering Alumni Association. What makes him a good teacher? “I try to lower the perceived wall between ‘faculty’ and ‘student’,” he says, “and to

be friendly and helpful as my teachers and industry mentors have been. I don’t make students follow a lot of rules, and instead, encourage them to focus on the engineering.”

For Vallancourt, the reward of teaching is a two-way street. “Chatting with students is a blast. I’ve met some really cool people who might not normally hang out with an old white-haired guy.”

Servedio and Vallancourt join five Engineering professors who have won the Presidential Teaching Award in the past. This list of outstanding educa-tors includes Patricia Culligan, professor of civil engineering and engineering mechanics; George Deodatis, Santiago and Robertina Calatrava Family Professor of Civil Engineering and Engineering Mechanics; George Flynn, Higgins Profes-sor of Chemistry and professor of chemical engineering; Yannis Tsividis, Charles Batchelor Pro-fessor of Electrical Engineering; and the late Nicholas Turro, Wil-liam P. Schweitzer Professor of Chemistry and former Co-Chair of Chemical Engineering.

By Melanie A. Farmer

Making the Grade: Servedio, Vallancourt Win Presidential Teaching Awards

Technology pioneer henning schulzrinne is once again all over the Net.

Schulzrinne, who is Julian Clar-ence Levi Professor of Math-ematical Methods and Com-puter Science at the Engineering School, has been inducted into the Internet Hall of Fame as a leading innovator. Schulzrinne is being recognized for co-develop-ing the key protocols that enable Voice over Internet Protocol, known widely as VoIP, and other significant multimedia applica-tions, including real-time web streaming and the Session Initia-tion Protocol (SIP), which sets up and configures VoIP telephony communications in businesses.

Schulzrinne is one of 32 indi-viduals who have been selected for the Hall of Fame this year by the Internet Society. Visionaries, innovators, and leaders around the world—from engineers to entrepreneurs—gathered August 3rd in Berlin, Germany, where they were officially inducted and honored at the So-ciety’s annual awards ceremony.

Since December of 2011, Schul-zrinne has served as the chief technology officer for the U.S. Federal Communications Com-mission (FCC). In this role, he guides the FCC’s work on tech-nology and engineering issues, together with the FCC’s Office of Engineering and Technology. He advises on matters across the agency to ensure that FCC policies are driving technologi-cal innovation, including serving

as a resource to FCC Com-missioners, and is helping the FCC engage with technology experts outside the agency and promote technical excellence among agency staff.

At Columbia Engineering, Schul-zrinne is also a professor of elec-trical engineering and former chair of the computer science department. His research inter-ests include Internet multimedia systems, applied network engi-neering, quality of service, and performance evaluation.

Schulzrinne received his under-graduate degree in econom-ics and electrical engineering from the Darmstadt University of Technology, Germany; his master’s degree as a Fulbright Scholar from the University of Cincinnati, Ohio; and his PhD from the University of Mas-sachusetts in Amherst. Before joining Columbia in 1996, he was a member of the technical staff at AT&T Bell Laboratories, Murray Hill, and an associate department head at GMD-Fokus (Berlin). He is a fellow of the Institute for Electrical and Electronics Engineers and former member of the Internet Architecture Board.

By Melanie A. Farmer

henning Schulzrinne Inducted to Internet Hall of Fame

Julian Clarence Levi Professor of Mathematical Methods and Computer Science Henning Schulzrinne

Associate Professor of Computer Science Rocco Servedio

Henning Schulzrinne is being recognized for co-developing the key protocols that enable Voice over Internet Protocol, known widely as VoIP, and other significant multimedia applications

Professor eu-gene Agichtein received the Karen Sparck Jones Award, which recogniz-es advances in our understand-

ing of Information Retrieval and Natural Language Processing with significant experimental contributions. Eugene received his Ph.D. in Computer Science from Columbia University in 2005 and is now an associate professor at Emory University. He is a Sloan Research Fellow.

Professor Augustin Chaintreau and coauthors from Stony Brook, AT&T, and Telefonica won a Best Paper

Award at the ACM/USENIX IMC conference held in Barcelona. The conference is about Inter-net measurement and although their contribution was submit-ted as a short paper, it received the highest rank among 178 submissions. In the paper titled “Follow the Money: Under-standing Economics of Online Aggregation and Advertising”, the authors propose to explore the hidden side of the web: the one where the money is made! While the web quickly evolves to accommodate new ad-based revenues and social services, the mobile social lab at Co-lumbia is working with various groups on making the econom-ics of Internet content and your data more transparent.

Professor steve Feiner received the 2014 Virtual Reality Career Award of the IEEE Com-puter Society Visualization and

Graphics Technical Committee. The citation reads, “In recogni-tion of his lifetime contribu-tions to augmented reality and virtual reality, including seminal research on mobile augmented reality, automated design and

layout, and applications to task assistance and navigation.”

Professor Roxa-na Geambasu received Honor-able Mention for Dennis M. Ritchie Doctoral Dissertation Award, which

recognizes research in software systems and to encourage the creativity that Dennis Ritchie embodied, providing a reminder of Ritchie’s legacy and what a difference one person can make in the field of software systems research. Roxana re-ceived the first honorable men-tion for the inaugural Ritchie Award competition!

CS PhD stu-dents kyung hwa kim, Jong Yul kim and EE PhD student hyunwoo nam won first place in the Juni-per/Comcast Software-De-fined Network (SDN) workshop and competition recently held at the Juniper Networks OpenLab facility in Bridgewater, NJ. Students were challenged to develop solutions using Juniper’s Junos Space Platform to improve net-work utilization and quality of user experience under dy-namic network conditions. The Columbia team, whose faculty advisor is Pro-fessor henning

schulzrinne, presented an application-aware SDN solution. They won for an excellent visu-alization tool to monitor network status, a differentiated routing algorithm that takes applica-

tion needs into account, and a compelling business model for service providers.

Professor Vishal Misra receives U. Mass Amherst Outstanding Alumni Award. In celebration of his significant

achievements and inspiring suc-cesses in his early career, Vishal will be awarded the 2014 Col-lege of Engineering Outstand-ing Junior Alumni Award. The citation reads: “for exemplary accomplishments, epitomiz-ing the potential of a U. Mass Amherst College of Engineering education.” The award will be presented on September 26, 2014 at U. Mass Amherst during Homecoming Weekend 2014.

CS Research Scientist Anargyros papageorgiou and Professor Joseph Traub’s publication “Measures of quantum computing speedup” was designated as Editors’ Sugges-tion in Physical Review A. The journal also pub-lishes a list of a small number of Physical Re-

view A papers that the editors and referees find of particular interest, importance, or clar-ity. These Editors’ Suggestion papers are listed prominently on http://pra.aps.org/ and marked with a special icon in the print and online Tables of Contents and in online searches. The paper introduces the concept of strong quantum speedup. It is shown that approximating the ground-state energy of an in-stance of the time-independent Schrodinger equation with d degrees of freedom and d large enjoys strong exponential quan-tum speedup. It can be easily solved on a quantum computer.

Some researchers in QMA theory believe that quantum computation is not effective for eigenvalue problems. One of the goals of this paper is to explain this dissonance.

Professor Roc-co servedio was awarded a three-year NSF grant to study learning and testing probabil-ity distributions.

A long and successful line of work in theoretical computer science has focused on under-standing the ability of compu-tationally efficient algorithms to learn and test membership in various classes of Boolean func-tions. The proposal of Rocco advocates an analogous focus on developing efficient algo-rithms for learning and testing natural and important classes of probability distributions over extremely large domains. The research is motivated by the ever-increasing availability of large amounts of raw unlabeled data from a wide range of prob-lem domains across the natural and social sciences. Efficient algorithms for these learn-ing and testing problems can provide useful modeling tools in data-rich environments and may serve as a theoretically ground-ed “computational substrate” on which large-scale machine learning applications for real-world unsupervised learning problems can be developed.

evangelia sitaridi was Awarded IBM PhD Fellowship, an intensely competitive worldwide pro-gram that hon-ors exceptional Ph.D. students who have an in-terest in solving problems that are important to IBM and fundamental to innovation in many academic

CS@CU SPRING 2014 11

disciplines and areas of study. Fellows are awarded a stipend for one academic year, and can apply for renewals for a total of up to three years. Eva Sitaridi is currently working with her advi-sor Professor kenneth Ross on database query process-ing algorithms using graphics processors.

Professor Venkat Venka-tasubramanian and coauthors won the Best Poster award at the 23rd Annual Meeting

of the European Symposium on Computer Aided Process Engineering (ESCAPE 23), held in Lappeenranta, Finland. This is a prestigious gathering of pro-cess systems engineers in the world, held annually at some European city. A committee of the conference organizers evaluate all the poster papers and announce the best poster paper award at the concluding session of the conference. This time, there were more than 100 posters presented from all over the world. The paper titled “In-telligent Alarm Systems applied to Continuous Pharmaceutical Manufacturing” discusses the development of a model-based intelligent control system that effectively integrates different models of process knowledge. This system provides timely guidance to the operator regard-ing the detection and diagnosis of exceptional events along with relevant mitigation strategies, with the goal of avoiding emer-gency shutdowns.

Professor Vladi-mir Vapnik was awarded the 2013 C&C Prize bestowed by NEC Founda-tion, the founda-tion arm of IT

and networking giant NEC Corp. He is recognized for contribu-tions to establishing Statistical Learning Theory and for the invention of high-performance and practical learning algo-

rithms. His breakthroughs were significant, and he has made valuable contributions to the development of machine learn-ing technology and the expan-sion of its application field, cited NEC C&C. Established in 1985, the C&C Prize is awarded to dis-tinguished individuals for their pioneering contributions related to the integration of computers and communications technolo-gies and the social impact of developments in these fields.

Dr. Adam waksman and CS Junior Mat-thew suozzo won a Best Student Paper Award at the 20th ACM SIG-SAC Conference on Computer and Communi-cations Security, held at Berlin, Germany in November 2013. The paper “FAN-CI: Identifica-tion of Stealthy Malicious Logic Using Boolean Functional Analysis” was co-authored with their advi-sor Professor simha sethu-madhavan,

which describes a method for detecting backdoors in hard-ware circuits before the design is taped-out and sent to the market. This is the first purely static analysis technique for de-tecting backdoors in hardware. Only three of 530 submissions received this honor.

Professor Junfeng Yang won Google Faculty Research Award. Mobile and wearable devices offer unprecedented

convenience to our lives, yet they also bring new threads. Studies have shown that the apps running on these devices

are often plagued with pro-gramming errors that degrade user experience. Worse, apps are sometimes infected with malware, which can steal users’ private information, install key loggers, fake mobile payments, etc. Junfeng and his team will use the funding from Google to build AppDoctor, a new system for detecting programming errors and malware in Android apps, benefiting every Android user.

A paper by Dr. lisa wu, EE student Ray-mond Barker, Professor Martha kim, and Professor kenneth Ross was named IEEE Micro “Top Picks from the Computer Archi-tecture Confer-ences” in its May/June 2014 issue. This issue collects some of this year’s most significant research papers in computer ar-chitecture based on novelty and long-term impact. Any computer archi-tecture paper

(not a combination of papers) published in the top confer-ences of 2013 is eligible. Top Picks will attempt to recognize those significant and insightful papers that have the potential to influence the work of computer architects for years to come. The paper that was recognized appeared in the International Symposium on Computer Archi-tecture (ISCA 13) and is entitled “Navigating Big Data with High-Throughput, Energy-Efficient Data Partitioning.”

Two Columbia publications were selected as “Top Picks in Computer Architecture” in the July/August 2013 issue of IEEE Micro. The issue features eleven of the year’s most significant research papers in computer ar-

chitecture based on novelty and long-term impact. The first paper is titled “Collection, Analysis, and Uses of Parallel Block Vectors.” Authored by PhD student Melanie kambadur, undergradu-ate kui Tang, and Professor Martha kim,

this research establishes a novel perspective from which to reason about the correctness and performance of parallel soft-ware. In addition, it describes the design and implementation of an open source tool that automatically instruments an program to gather the neces-sary runtime information.

The second paper is titled “A Quantitative, Experimental Approach to Measuring Processor Side-Channel Security.” The

authors are Dr. John Demme, MS student Robert Martin, Dr. Adam waksman and Profes-sor simha sethumadhavan. This paper describes quantita-tive method to identify bad hardware design decisions that weaken security. The methodol-ogy can be used in the early processor design stages when security vulnerabilities can be easily fixed. The paper marks the beginning of a quantitative approach to securing computer architectures.

Department News & Awards

10 CS@CU SPRING 2014

Evangelia Sitaridi

Professor Kenneth Ross

Melanie Kambadur

Kui Tang

Kyung Hwa Kim

Jong Yul Kim

Hyunwoo Nam

Professor Henning Schulzrinne

Anargyros Papageorgiou

Professor Joseph Traub

Dr. Adam Waksman

Matthew Suozzo

Professor Simha Sethumadhavan

Dr. John Demme

Dr. Lisa Wu

Raymond Barker

Professor Martha Kim

NEWSLETTER OF THE DEPARTMENT OF COMPUTER SCIENCE AT COLUMBIA UNIVERSITY WWW.CS.COLUMBIA.EDU

Columbia University in the City of New YorkDepartment of Computer Science1214 Amsterdam AvenueMailcode: 0401New york, Ny 10027-7003

ADDRESS SERVICE REQUESTED

NoN-PRofIT oRG.

U.S. PoSTAGE

PAID

NEW yoRK, Ny

PERMIT 3593